AUTHOR=Egger Roman , Yu Joanne TITLE=A Topic Modeling Comparison Between LDA, NMF, Top2Vec, and BERTopic to Demystify Twitter Posts JOURNAL=Frontiers in Sociology VOLUME=Volume 7 - 2022 YEAR=2022 URL=https://www.frontiersin.org/journals/sociology/articles/10.3389/fsoc.2022.886498 DOI=10.3389/fsoc.2022.886498 ISSN=2297-7775 ABSTRACT=Seeing that travel restrictions have begun to ease up, thanks to increasing vaccination rates across the globe, understanding the public’s concerns and opinions regarding traveling during the pandemic has the potential to foster the kickstart of tourism recovery. One such approach is through disentangling the wording of an enormous amount of social media posts. However, although social media has paved a new way to gain insights, their short and unstructured nature often leads to methodological challenges in data collection as well as analysis. In order to bridge the developing field of computational science and empirical social research, this study aims to evaluate the performance of four topic modeling techniques. By extracting Twitter posts related to COVID-19 and travel, a comparison can be made between latent Dirichlet allocation (LDA), non-negative matrix factorization (NMF), Top2Vec, and BERTopic. Thus, in view of digital media and contemporary tourism characteristics, this research assesses the performance of different algorithms and discusses their strengths and weaknesses in a sociological context. Based on certain details during the analytical procedures and on quality issues, this research sheds light on the efficacy of using BERTopic and NMF to analyze Twitter data.